This study focuses on the problem of tuberculosis (TB), the most important opportunistic infection in many HIV risk populations, including the homeless. When a susceptible person inhales a TB bacterium, TB infection is established. Immunocompromised persons like the HIV-infected an progress rapidly from asymptomatic TB infection to active TB disease. Active disease is life-threatening and infectious, spreading bacteria through the air to others. The social conditions of homelessness, such as group shelters, maximize contact between those at risk for HIV and those with active TB disease. We showed recently that 9% of homeless adults in San Francisco are HIV-infected and 32% are TB-infected. In order to reduce the risk of TB infection to HIV-infected persons, the homeless, and the community at large, it is essential to minimize the number of people with active TB disease. A standard way to do this is to block the preogression from TB infection to active TB disease by six months of prophylaxis with isomiazid (TNH). However this TB prevention strategy is not used for the homeless because of poor adherence to the long course of INH therapy. The proposed study consists first of screening homeless and unstably housed adults to identify TB and HIV infected persons, and second of a clinical trial of methods to improve adherence to INH preventive therapy. We will conduct a three-arm randomized controlled trial of methods to give six months of INH therapy to TB-infected homeless persons. We will compare the use of (1) community-recruited "peer health advisors" and (2) direct cash incentives to (3) usual care provided by the TR clinic. We will determine the extent to which adherence to TB prevention in the trial is predicted by (a) sociodemographic characteristics, (b) regimen knowledge and TB clinical environment, (c) psychological and behavioral factors, and (d) characteristics of the social environment of homelessness. As a result of screening for the trial, we will be able to estimate new infection rates for both HIV and TB in the homeless. We will identify high risk groups for both infections and examine specific socioeconomic and behavioral risk factors. We will examine heterosexual transmission of HIV and the effects of needle exchange use on HIV seroconversion.